Prompt & Target Platform

100% in-browser evaluation
Default training enabled; conversation profile linked to Meta ad graph.
Quick Test Scenarios 0 chars
Ready. Paste text or choose a scenario.

Exposure Risk Assessment

Default Model Training
YES Enabled by Default
Your inputs are incorporated into future weight updates and synthetic datasets.
Retroactive Opt-Out
NO Irreversible
Opting out today does not purge previously trained model parameters or caches.
Ad Profile Linkage
FORCED Linked
Conversations are indexed to build ad interest profiles across company services.
Human Review / Log Retention
180+ Days
De-identified transcripts sampled by human annotators with multi-year archives.
Detected Sensitive Tokens (0) Scanning for credentials, PII & health data...
Sanitized Safe-to-Send Version 
No sanitized output yet. Paste text or run redaction.

Why AI Chat Privacy Policies Matter

1. Why Opt-Outs Aren't Retroactive

Once your prompt data passes through gradient descent into foundation model weights, "machine unlearning" is mathematically infeasible without retraining the entire model at immense expense. Deleting your chat history merely deletes UI log pointers, not the baked-in parameter associations.

2. The Meta AI Ad Targeting Mandate

Unlike enterprise LLM providers, Meta explicitly incorporates consumer AI chats into user behavioral graphs. When you discuss medical symptoms, vacation plans, or purchase intent in WhatsApp or Instagram AI chats, those topics feed algorithmic ad targeting across its network.

3. Consumer Free Tiers vs. API Endpoints

Most AI companies subsidize free consumer interfaces by retaining prompt logs for fine-tuning. Conversely, commercial APIs (such as OpenAI API or Anthropic API) enforce zero-day retention and contractually guarantee no model training on customer payloads.

Enjoy this tool? Build your own with Super